I am trying to find the cumulative sum of x=c(1,2,3,4,5) based on moving window of y=c(1,1,2,3,2). I tried the following code based on rollapply in zoo package:

rollapply(c(1,2,3,4,5),c(1,1,2,3,2),cumsum,align="right",partial=T)

My expected answer is 1,2,5,9,9 (because I am cumulating from the left side) but instead I get the below 4x3 table:

How can I correct the code to get the desired result?

I think you should use sum instead of cumsum

> zoo::rollapply(c(1, 2, 3, 4, 5), c(1, 1, 2, 3, 2), sum, align = "right", partial = T)
[1] 1 2 5 9 9

and you can use mapply as well to achieve the same output

mapply(\(a, b) sum(x[b + 1 - a:1]), y, seq_along(y))

Benchmark

set.seed(0)
x <- seq.int(1e4)
y <- sort(sample(1:10, length(x), TRUE))

library(zoo)
microbenchmark(
  rollapply = rollapply(x, y, sum, align = "right", partial = T),
  mapply = mapply(\(a, b) sum(x[b + 1 - a:1]), y, seq_along(y)),
  check = "identical",
  times = 10L,
  unit = "relative"
)

shows rollapply is not efficient as mapply

Unit: relative
      expr      min      lq     mean  median       uq      max neval
 rollapply 9.362684 9.32441 9.340693 9.31007 9.985161 9.342603    10
    mapply 1.000000 1.00000 1.000000 1.00000 1.000000 1.000000    10